On the surface, this reads as a routine pricing update. We think it’s bigger than that.

For decades, marketing and sales operations have struggled to put a clean price tag on knowledge work. We all know roughly what an MQL “costs” — on paper. In reality, that number rarely captures the SDR’s salary, their manager’s time, the CRM licenses, the training, the turnover cost when they leave in 8.5 months. Customer support is worse. Productivity in a support queue depends on ticket complexity, team morale, tooling, and a dozen other variables nobody wants to (or can) model in a spreadsheet. Executives have been making staffing calls based on fuzzy averages for as long as these functions have existed.

Now, a qualified lead has a price: one dollar. A resolved ticket has a price: fifty cents.

Whether those numbers are accurate for your business is a separate conversation, and the definitions of “resolved” and “qualified” deserve real scrutiny before anyone signs a contract. But the psychological shift at the executive level is enormous. Cheapness is part of it — fifty cents a ticket is hard to argue with. The bigger thing is legibility.

AI work shows up on the P&L with a clean unit cost, no footnote required, and you can forecast against it like you forecast cloud spend. Predictability beats cheapness, nine times out of ten.

The interesting question is where this goes next. Image generation already has a per-output price. Video generation does too. Today, those numbers sit in the “cost of experimentation” column on most budgets, often tucked quietly into an agency invoice somewhere. We’re probably one or two product cycles away from somebody — Canva, Adobe, a challenger we haven’t heard of yet — charging per approved creative asset, per campaign variant that survived brand review. When that happens, the last great unpriced domain in marketing finally gets a price tag.

The conversations we’re having today about AI in support and sales will feel like the warm-up act.

— Torsten and Peter

Share with a colleague. Earn paid membership months.

Refer a friend

Everything we know about AI marketing. Free.

Courses, workshops, frameworks, tool reviews, and implementation guides — all free, all on-demand. Built for senior marketers who don’t have time for theory.

→ Start Browsing The Hub

3 AI Marketing Tools To Try Today

Brandwell

→ Try it here

Brandwell is an AI long-form content engine that generates research-backed, SEO-optimized blog posts designed to bypass AI detection. Use it to scale organic content production without sacrificing quality or search rankability.

SitesGPT

→ Try it here

SitesGPT uses AI to generate fully responsive websites in minutes by analyzing your business information and applying industry best practices. Use it to spin up professional landing pages and microsites without design or dev resources.

Birch

→ Try it here

Birch is an AI-powered ad automation platform that manages, optimizes, and scales paid campaigns across Meta, Google, TikTok, and Snap with real-time KPI monitoring. Use it to stop babysitting ad accounts and let automation handle bid adjustments, pausing, and scaling.

Paid members, your toolkit and this month’s report are below.

Read more

That is about to flip in the most ironic way.

The next generation getting accused of hogging the resources and blocking the people below them is going to be millennials themselves — still in their jobs, still in the promotions, just AI-fluent enough to stay put. And before you file this under "HR problem," note that the younger generation being squeezed out is also your future buyers and your future hires. How they feel about AI at work is how they'll feel about AI in your marketing, and your marketing department.

You can see this in two reports published last week, and they make much more sense read together. Gallup’s new Voices of Gen Z study surveyed more than 1,500 14-29-year-olds and found Gen Z’s relationship with AI souring fast. Excitement and hope are down sharply year over year. Anger is up. Nearly half of employed Gen Zers now say the risks of AI at work outweigh the benefits. And this isn’t skepticism that fades with exposure. Even daily users feel less positive than they did twelve months ago.

Writer’s AI Adoption in the Enterprise report, surveying 2,400 executives and employees, shows the view from the other end of the org chart. 60% of companies plan to lay off employees who can’t or won’t use AI. The reaction is that 44% of Gen Z employees actively sabotage their company’s AI rollout. Not just passively resisting but feeding proprietary data into unapproved tools, tampering with outputs to make AI look ineffective, or even refusing to use the tech on principle.

Executives see AI as leverage for headcount cuts. Workers see AI as a job threat and either disengage or sabotage the rollout. Executives take the sabotage as proof the workforce needs restructuring. Repeat. The resistance is accelerating the exact outcome that sparked it.

Which is how millennials end up as the unlikely winners here, not through enthusiasm, but through position. They’ve aged out of the entry-level roles getting automated first, aged into enough seniority that companies hesitate to cut them wholesale, and have just enough patience with the tools to clear the bar.

Survivorship is how you become the next generation everyone quietly resents.

But back to the marketing angle: Gen Z isn’t just your entering workforce, they’re your next decade of buyers. Gallup found they trust work produced without AI more than twice as much as AI-assisted work. Every AI-generated ad, every chatbot, every personalized email sequence is being received by an audience that is, on average, getting more suspicious of the category, not less. The assumption that exposure breeds acceptance looks shaky at best.

Which leaves an awkward question on the table for us building AI-forward marketing functions: are we building leverage, or are we building the thing our customers will eventually resent us for? The honest answer is that we don't know yet.

Just maybe, AI in marketing is about to become what outsourced call centers were in 2005 — a cost-saving move that quietly damaged the brands that overused it and quietly advantaged the ones that held the line. The winners of the last cost-cutting wave weren't the companies that outsourced best.

They were the ones that didn't.

— Torsten and Peter

Share with a colleague. Earn paid membership months.

Refer a friend

Everything we know about AI marketing. Free.

Courses, workshops, frameworks, tool reviews, and implementation guides — all free, all on-demand. Built for senior marketers who don’t have time for theory.

→ Start Browsing The Hub

3 AI Marketing Tools To Try Today

HeyGen

→ Try it here

HeyGen creates lifelike AI avatar videos with voice cloning and multilingual support, generating professional video from text prompts. Use it for personalized outreach videos and multilingual explainers with avatar quality a step above most competitors.

Nuelink

→ Try it here

Nuelink is an AI-assisted social media automation tool that schedules, recycles, and cross-posts content with smart captions. Use it to keep evergreen content running on autopilot across all your channels as a solopreneur or small team.

Tella

→ Try it here

Tella uses AI to record and auto-edit screen recordings, cutting mistakes, silences, and filler words for polished video output. Use it to save significant post-production time on product demos and walkthroughs with intelligent auto-editing.

Paid members, your toolkit and this month’s report are below.

Read more

This workshop is part of our new Teamless initiative. The idea is simple: the teams that figure out agents now will need fewer people to do more work. That cuts both ways — as a leader managing headcount, and as a professional managing your own career.

The recording is now on YouTube. Watch it here.

If you’d rather read the summary first, here are 10 things you’ll learn from watching it.

1. Agentic companies have six building blocks. Company, teams, agents, projects, tasks, and skills. Understanding how these fit together is the prerequisite for building anything that actually runs. Miss one, and the whole system breaks down.

2. You don’t need to redesign your whole company. You can introduce agentic workflows into a single team — even a sub-team. Start small, prove it works, then expand. The architecture supports it from day one.

3. The company layer is context for agents, not humans. Your brand, your products, your tone of voice, your goals — all of this needs to be written down so agents can reference it when they don’t have enough context. Think of it as the onboarding document for every agent you’ll ever hire.

4. Agents are roles, not skills. An agent has a job description, reporting relationships, and a defined role — just like a human colleague. The mistake most people make is baking the technical how-to into the agent. That’s what skills are for.

5. Skills are reusable. Tasks are not. A skill is “how to navigate the Meta Ads dashboard.” A task is “set up campaign X with this budget and this audience.” Skills can be shared across multiple agents. Tasks are one-time instructions with a clear input, output, and deliverable.

6. The real power is in separating role from capability. Your media buyer agent and your reporting agent can both use the same Meta API skill. You don’t duplicate the knowledge — you share it. That’s the compounding efficiency that makes agentic teams fundamentally different from human teams.

Read more

About the guest

is a strategy and management consultant based in the UK, specialising in research and higher education. He works with universities, research funders, and corporate actors both nationally and internationally — advising clients on strategy, operations, and complex governance challenges. He is the author of on Substack, where he writes about research methodology, consulting practice, and how to think more clearly in a world flooded with AI-generated output. He also builds AI-powered tools for consultants, including a question methodology checker that flags bias and recall errors in research instruments.

Connect with Andrea on The Art of Asking Questions on Substack and LinkedIn.

About the host

co-founded , the daily intelligence platform for senior marketing leaders. Peter has been serving as a CMO and marketing leader and consultant to high-growth B2B scaleups for the past 10+ years. He has a background in advertising, working with Fortune 500 brands.

Connect with Peter on LinkedIn or read his newsletter.

Top 10 Takeaways

1. AI hasn’t replaced consultants — it’s replaced the formula-based ones. — If your consulting model was recycling the same advice for every client, AI can do that. If it were based on judgment, context, and human understanding, you’re still fine.

2. You can generate a million pages of ideas in minutes. That’s not the hard part. — The hard part is knowing which ideas are right for this specific client, in this specific context, right now. AI can’t do that.

3. The art of asking questions is irreplaceable. — Question order, framing, and bias aren’t mechanical problems. They’re deeply human ones. Getting them wrong produces biased data — and biased data produces six-figure strategic mistakes.

4. Use one LLM to produce, another to validate. — Run your AI-generated report through a second model to fact-check claims, sources, and logic. The disagreements between the two are your red flags. This can halve your validation time.

5. Watch where two AIs disagree — those are the high-risk areas. — When models conflict on a source or claim, that’s your signal to check manually. Most of the rest is routine click-through.

6. AI made consultants more efficient, but made their time worse. — Producing more in less time sounds like a win. But replacing deep reading with tedious validation is a real trade-off that most efficiency narratives ignore.

7. Transparency about AI use isn’t optional — it’s a reputation issue. — In consultancy, if a client discovers you used AI without disclosing it and the output was wrong, your credibility is gone. Disclose at kickoff. Confirm they’re comfortable.

8. The story is the last thing you can outsource. — Data collection, extraction, and chart generation — all automatable. The narrative that makes findings land with a specific audience, at this moment, in this context, is still entirely human.

9. AI is creating a hidden junior pipeline problem. — Senior consultants doing their own research with AI means junior colleagues never get the work that used to train them. The talent pipeline is quietly breaking.

10. Solo consultants now compete directly with big agencies on price and speed. — Two or three experienced practitioners with good AI workflows can be faster, leaner, and more competitive on pricing than a 40-person agency. That dynamic is already shifting the market.

5 Things Worth a CMO’s Attention

1. Why consulting isn’t dying — and which consultants actually should worry

The narrative that AI will kill consulting is too blunt to be useful. Andrea’s distinction is sharper and more actionable: the consultants at risk are the ones whose model was always formula-based — the same framework, the same slide deck, the same recommendations, slightly customized for each client. AI can replicate that. It already does.

The consultants who are genuinely safe are the ones whose value was always in judgment — the ability to read a specific organization’s dynamics, politics, and constraints, and give advice that fits that exact context. That’s not something you can extract into a prompt. The relevant variables aren’t in any document. They exist in conversations, in the silences, in the things people don’t say.

For marketing leaders who use consultants — or who are consultants themselves — this reframes vendor selection. The question to ask any external partner is no longer “what’s your methodology?” It’s “what do you bring that I can’t get from a good prompt?” If they can’t answer that clearly, the model is probably formula-based. And formula-based is exactly what AI is eating.

2. The art of asking questions — why bad research design costs more than bad data

Andrea’s newsletter title isn’t just branding. It’s a thesis. The single most important skill in research — whether you’re a consultant running stakeholder interviews or a marketer designing a customer survey — is asking questions that don’t contaminate the answers.

The example he gives is instructive: “How often do you go to a restaurant?” is a deeply flawed question because people can’t recall frequency accurately without a time period. The data you collect is unreliable. And if that data is backing a strategic decision — a positioning call, a campaign budget, a new product bet — the flawed research propagates all the way to the outcome.

AI makes this problem more acute, not less. AI can generate a hundred interview questions in seconds. It cannot tell you whether those questions introduce bias, whether the ordering creates a leading effect, or whether they’re structured in a way that distorts recall. That evaluation requires methodological expertise and judgment. Without it, you’re just producing bad research faster.

The practical takeaway for marketing teams: use AI to generate question candidates, then apply human judgment to evaluate them for bias, framing, and order. And if your team doesn’t have that methodological expertise in-house, Andrea’s question-checker tool on his Substack is a good starting point.

3. The dual-LLM validation trick — and how to use disagreement as a signal

One of the most practically useful things that came out of this conversation: the technique of using one LLM to produce content and a second LLM to validate it.

The workflow is straightforward. You generate a report, analysis, or research summary with your primary model. You then pass the full output to a second model — or an agentic tool like Claude’s computer use — with a clear instruction to fact-check every factual claim, verify every source, and flag any logical inconsistencies. You review the flagged items manually.

Andrea and Peter both use versions of this workflow, and both have found that it significantly reduces validation time without eliminating it entirely. The key insight Andrea adds is to pay special attention to the areas where the two models disagree. When one model makes a claim and the second model questions it, that’s your red flag. That’s where the risk of error is highest. The rest — the claims both models agree on — you can move through much faster.

This doesn’t solve the validation problem entirely. It changes its character. You’re no longer reading everything in depth. You’re triage-reviewing a flagged list. That’s faster. But Andrea’s honest about the trade-off: the deep reading was more intellectually engaging. The triage is tedious. Efficiency doesn’t mean the experience of work has improved.

4. The storytelling gap — why the narrative is what you actually get paid for

One of the clearest arguments in this conversation is about where AI bottoms out in the research-to-reporting workflow.

Data collection: largely automated. Extraction and structuring: highly automatable. Visualization and charting: increasingly automatable. The executive summary that makes the findings land with a specific C-suite audience, in a specific organizational moment, in a way that moves them to act: still entirely human.

Andrea has built Claude prompts that generate multiple narrative outline options from a set of research findings — different structural approaches to telling the same story. That’s useful as a starting point. But choosing which narrative arc will resonate with a particular audience, at this particular moment, given what you know about the politics and priorities of the room — that requires the kind of instinct that only comes from experience.

For CMOs presenting to boards or senior leadership, this is the point worth internalizing. AI can help you structure the data. It can suggest frameworks. It can write draft copy. But the story — the one that actually lands, that actually changes something — requires you to bring judgment about your specific audience that no model can replicate. The story is not a deliverable you can outsource. It’s the thing you’re actually being paid for.

5. The junior pipeline problem — a slow-burn talent crisis hiding inside AI efficiency gains

This is the issue neither productivity dashboards nor efficiency reports will surface, but Andrea names it clearly: as senior consultants and senior marketers use AI to do their own research and production work, junior colleagues stop getting the assignments that used to develop them.

The old model: a senior consultant scopes a project, briefs a junior colleague on the research task, reviews their output, gives feedback, and iterates. The junior colleague learns by doing. The senior colleague invests time in that exchange.

The new model: the senior consultant does the research with AI. It’s faster. There’s no time for the exchange. The junior colleague never gets the assignment.

This has an obvious short-term efficiency win. It has a serious long-term talent problem. The people who will be mid-level and senior in five years are the junior consultants and junior marketers who are supposed to be developing right now. If AI is absorbing the work that develops them, the pipeline thins — and eventually the organization has no one to promote.

For marketing leaders, this is worth explicitly building into AI adoption planning. Efficiency gains at the senior level that come at the cost of junior development are not free. Someone is paying for them — it’s just a cost that won’t show up on a balance sheet for three to five years.

But quantify it? Predict it? That was for people who didn’t understand The Craft.

Well. About that.

Two weeks ago, Meta open-sourced TRIBE v2, a foundation model trained on over 1,000 hours of fMRI recordings from 720 volunteers watching movies, listening to podcasts, and reading text. Feed it a video and it predicts, second by second, how 70,000 points in a human brain would light up in response. Meta’s own research shows TRIBE v2’s predictions are often more representative of the typical neural response than any single actual fMRI scan, because real scans are noisy and the model strips the noise out. It’s a digital mirror of attention itself.

You can try it, today. It requires minimal technical knowledge, not much more than knowing where to click. And it’s free.

A marketer on X ran an already-edited UGC video through the model, saw exactly where brain engagement spiked and where it flatlined, moved the high-engagement moments earlier, cut the dead zones, and reposted. The re-edit did 221,100 views. Previous videos from the same creator were doing a small fraction of that. One pass, minimal effort. A file of predicted neural activity acting as the world’s most honest focus group.

This is the part where a think-piece like this usually pivots to “of course, the human touch still matters.” We’re not going to do that, because we don’t think it’s true for much longer.

What TRIBE v2 really represents is the moment when creativity stops being a mystery and starts being a measurable signal with a loss function. You can A/B test against a brain now.

And once you can do that, somebody will build the pipeline that does it automatically—twenty variations in, one brain-optimized winner out.

The logical endpoint is a world where the threshold of attention becomes impossibly high. Every piece of content you encounter will have been pre-screened against a model of your own neurology. The dull bits, the setup beats, the moments a human editor would have left in because they felt right—gone, because the simulated cortex didn’t light up.

Everything that survives will be optimized brainrot in the literal sense: content engineered to keep predicted neural response elevated from frame one to frame last. Good luck launching a slow-burn campaign into that feed.

We’re honestly not sure how to feel about this. It’s the most interesting thing we’ve read in weeks, and also a reasonable preview of a future we’re not particularly looking forward to. The craft we romanticized is about to get a measurement layer, and measurement layers have a way of collapsing taste into whatever the metric rewards.

But, there is a silver lining.

The skill that still isn’t quantifiable is deciding what to make in the first place: the strategic judgment about what deserves to exist.

Hold onto that one tightly.

It’s the last part of the job the brain scanner doesn’t have an opinion on yet.

— Torsten and Peter

Share with a colleague. Earn paid membership months.

Refer a friend

3 AI Marketing Tools To Try Today

Crayo

→ Try it here

Crayo is an AI short-form video creator that generates clips with auto-captions, backgrounds, and music optimized for TikTok, Reels, and Shorts. Use it to produce scroll-stopping vertical video at scale without touching a timeline editor.

Frase

→ Try it here

Frase combines AI-powered SERP analysis, content briefs, and writing in a single tab, streamlining the entire SEO content workflow. Use it for a seamless brief-to-draft workflow that’s hard to beat on value in the SEO content category.

Murf AI

→ Try it here

Murf AI generates 300+ natural-sounding AI voices in 33 languages with customizable tone, pacing, and emotion from text input. Use it to skip the time and cost of booking voice talent for ads, explainer videos, and multilingual campaigns.

Paid members, your toolkit and this month’s report are below.

Read more

90.3% of companies report using AI agents.

Only 6.3% have fully integrated them into their marketing stack.

We’ve seen this trend before with genAI. And it just amplifies if we look at the preferences for adoption drivers: enterprises prefer custom-built solutions, while SMBs prefer integration platforms like Make, Zapier, etc. Which basically means, they “hook things together DIY” in plain English, let’s be honest here.

We love these studies because they show trends. But everyone should think twice before accepting the numbers as is. As said, we’ve seen this trend before with genAI. Out of 10 companies, 8 said they used AI in their marketing. But when you look deeper, the “used AI in their marketing” meant a GPT Pro subscription only. At best, because most only chose the free version.

Get 10% OFF forever on AI-Ready CMO membership.

Get 10% off forever

Don’t know about you, but we wouldn’t call this AI adoption. Maybe AI curiosity, yes, but it’s miles away from AI integration.

Now we have the same trend for AI agents. Out of 10 companies, 9 claim to have agents. Meanwhile, it might mean that they mistakenly call automation agents. Just look at the example tools, n8n, Zapier, and Make, all of which are more like automation tools, less like agentic workflows. Custom-built for enterprise? That’s the same, hand-built for them by expensive Salesforce consultants, probably.

One thing that is clear from studies like this: organizations have structural problems. Most organizations are layering AI onto systems built without agents in mind. The agent needs to know the rules, the customer context, and the appropriate action. In real time, across multiple systems simultaneously. That’s a structural problem to solve, not a knowledge challenge to wing.

It’s architecture. It’s scaffolding. We covered exactly how to build this scaffolding in our free workshop. Give it a watch:

TLDR: solve the structural problem first by reorganizing your workflows with agents in mind. Then deploy the agents to scale your workflows. Agentic workflows are not just automations. They are your new team members, doing your work 0-24, acting autonomously and automatically.

— Torsten and Peter

Share with a colleague. Earn paid membership months.

Refer a friend

3 AI Marketing Tools To Try Today

Make

→ Try it here

Make is an AI-enhanced visual automation platform that connects your entire marketing stack through drag-and-drop workflows with conditional logic and API calls. Use it when your automations need branching paths, error handling, and complex multi-step sequences beyond simple triggers.

n8n

→ Try it here

n8n is an open-source workflow automation platform with 400+ integrations and AI capabilities for building complex business process automations. Use it as a fully customizable alternative to Zapier for marketing ops teams with specific integration needs.

Gumloop

→ Try it here

Gumloop is a no-code AI automation platform that lets you visually chain together AI models, web scrapers, and data tools to build complex workflows. Use it to automate repetitive processes, such as lead enrichment or content pipelines, without writing a single line of code.

Paid members, your toolkit and this month’s report are below.

Read more

We called this workshop part of our new Teamless initiative. The idea is simple: the teams that figure out agents now will need fewer people to do more work. That cuts both ways — as a leader managing headcount, and as a professional managing your own career.

The recording is now on YouTube. Watch it here.

If you’d rather read the summary first, here are 10 things you’ll learn from watching it.

1. Agents do work. Chatbots just respond. A chatbot answers questions. An agent opens your browser, reads competitor websites, writes a Word document, saves it to a folder, and moves to the next task — without you touching a thing.

2. Claude Cowork sits between Chat and Code. It’s Claude’s desktop agent. More capable than the web chat. Much friendlier than Claude Code, which is built for developers. If you’re a non-technical marketer, this is your entry point into agentic work.

3. Projects are just folders — and that’s a good thing. Cowork organizes work around projects. A project can be a campaign, a quarter, or your whole brand. Everything Claude learns and builds stays inside that project context and compounds over time.

4. One prompt. Five competitor websites. One Word doc. Torsten asked Claude to research five baby-wearing brands, extract key messages, voice, tone, positioning, and competitive gaps, and deliver it as a formatted document. It would have taken a junior marketer three days. It took Claude about ten minutes.

5. Agents get smarter the more you give them. After the competitive research, Claude built a tone-of-voice guide and a key messages document from scratch — based solely on what he had already created for the project. No new uploads. No re-briefing. That’s the compound knowledge effect.

6. You can build an email drip campaign in a single prompt. Using nothing but the brand context it had already built, Claude wrote a five-email welcome sequence with subject lines, preview text, and full copy. If you have HubSpot connected, it can upload those emails directly.

7. Agents can connect to your existing tools. HubSpot, Supermetrics, CMSs, analytics platforms — Claude Cowork has connectors for most of the stack you’re already running. Your agent doesn’t just write about your data. It can work with it directly.

8. Ad performance analysis in 20 minutes. No Excel formulas. Torsten uploaded a raw Meta Ads export. Claude wrote Python behind the scenes, ran cohort analysis across age groups and targeting approaches, and returned a formatted Excel sheet with color-coded analysis and a clear narrative on what was driving cost per conversion.

9. Skills turn one-off prompts into reusable workflows. Once Torsten had the ad analysis working, he saved it as a skill called “meta-ad-analysis.” Next time he runs the raw export, a single slash command will run the entire analysis automatically. Skills are repeatable work with zero repetition.

10. Scheduled tasks mean the work runs without you. Claude Cowork supports scheduled tasks. Torsten set the ad analysis to run every weekday at 4 pm, the interactive dashboard to generate at 4:30, and the report to email his manager at 4:40. The whole reporting chain runs automatically while his computer stays awake.

Watch the full workshop here.

If this is a conversation you want to have for your marketing team — a hands-on workshop on building agentic workflows for real marketing use cases — reply to this email. We’re exploring customized sessions for teams that want to move from AI chat to AI execution.

The March report is now available to Paid members. This is the version for everyone else.

We titled this one The Builders’ Quarter. Not because it’s an inspiring sentiment but because March produced a specific kind of evidence: the people and organizations that are building things with AI right now — actual tools, actual workflows, actual outputs — are separating from those who are still orienting themselves.

To be honest, time is running out. For most of you, it’s already too late. The gap is becoming visible. It will become measurable by Q3.

Three things drove that conclusion. They’re worth understanding separately before you see how they connect.

The content assumption cracked.

Two studies landed in March that most marketing leaders have yet to reconcile.

The New York Times published a blind writing test. Readers preferred AI-generated writing 54% of the time. Not a marginal result. A majority. The practical reading is not that AI writes better than humans — it’s that for most content formats, audiences are not making that distinction and would not change behavior if they did.

HBR published the opposite problem. Seven leading LLMs were tested across 30,000 simulations on fundamental strategic questions. Every model gave the same fashionable answer regardless of context. Differentiate rather than commoditize. Long-term over short-term. Always augment, never automate. The researchers called it “trendslop.” Additional prompting and context barely moved the result.

Refer a subscriber and unlock paid membership for free.

Refer a friend

Together, these two findings create a clear picture that most content strategies have not yet been priced in.

For commodity content — the daily post, the email sequence, the product page — authorship has no audience value. AI produces it at quality, at speed, at near-zero cost. The case for human production on these formats is effectively closed. For authority content — a named person taking a public position, a live conversation, a piece of analysis where the author’s reputation is on the line — authorship is the entire value proposition. You can disagree with a person. You cannot disagree with a synthesis.

The strategic implication is a question of reallocation. Where are your senior people currently spending hours on commodity formats? That time is better spent elsewhere, in formats where human presence actually changes what the audience does next.

Which brings us to where those authority formats actually exist in 2026.

The channels where human-to-human communication still operates at scale are contracting. Substack functions differently from social media because the distance between writer and reader is structurally shorter — comments, replies, a two-way contract rather than a broadcast. Live video retains human signal value for now; real-time presence is genuinely difficult to replicate convincingly. In-person is increasingly premium due to scarcity.

Forrester reports that a third of consumers choose offline over online brand experiences. Edelman puts social media trust at 42%. When Europol’s projection of 90% AI-generated online content materializes, physical presence will be the only channel with a verifiably human origin.

Social media, for most B2B brands, sits in an uncomfortable position: commodity content quality, produced at human cost, with a declining authorship premium. That is a resource allocation problem with a clear solution.

The org chart assumption cracked.

Austin Lau was Anthropic’s first growth marketing hire. For ten months, he ran paid search, paid social, app store optimization, email, and SEO at a company now valued at $380 billion. No engineering background. No team. He described problems in plain English, and Claude Code built him production tools. Ad creation time dropped from two hours to fifteen minutes. Creative output increased tenfold. He reviewed every output personally, because the judgment layer did not change — only the execution layer beneath it.

We wrote about this trend in our edition, The army of one:

The internet told this story as a marvel. We read it as a prototype.

The layoff data from Q1 confirms it. Block halved its staff while profitable. Microsoft cut 7,000 roles during a period of record earnings. Amazon’s CEO stated that as agent deployment scales, the company will need fewer people. What these organizations are removing is not aggregate headcount — it is the middle layer specifically. The coordination, oversight, and translation functions that middle management provided are being absorbed by AI. What remains is senior judgment at the top and automated execution beneath it.

We wrote about this trend in our weekend edition, The morning after the fire:

The university system is making the same calculation from the supply side. China’s Ministry of Education revoked more than 1,400 programs in 2024. In March, the Communication University of China suspended 16 undergraduate creative degrees — photography, visual communication, and translation. Jilin, East China Normal, and Tongji followed. These are not individual institutions making independent assessments. This is policy. Western universities will move more slowly, but the market logic is identical: a four-year program that trains students in a specific execution skill that AI now performs at scale is not a viable offering.

We also wrote about this in our daily edition here:

The org design question for a marketing leader in Q2 2026 is not whether this is happening. It is how far your organization’s response has progressed. Mapping your team’s work between judgment — strategy, positioning, creative direction, stakeholder alignment — and execution — content production, reporting, campaign management, routine analysis — gives you the starting point. Execution-heavy roles are build candidates. Judgment-heavy roles warrant investment. The organizations completing this analysis now will enter Q4 budget conversations with clarity rather than pressure.

This trend was an inspiration for our new approach, which we have detailed here:

The expectation assumption cracked.

A year ago, “I’m building my AI capabilities” was a forward-thinking professional position. In March 2026, it will become inadequate.

Board and executive conversations have moved to specificity. The question is no longer whether marketing leadership uses AI. It is what is running autonomously, what the results are, and whether those results are documented.

Currently, 6% of marketers have fully implemented AI in their workflows. 90% of the pressure for AI adoption comes directly from the C-suite. That gap — between organizational expectation and individual delivery — is where careers are most exposed right now.

Read more

About Patrice Poltzer

Patrice Poltzer spent seven years as a producer on NBC’s TODAY Show, where she had a front-row seat to the power — and devastation — of how a story, well told or badly told, could change the trajectory of a brand overnight. In 2016, she left to build her own storytelling agency in New York, working with brands like Netflix, Amazon, and Lululemon.

In 2024, she co-founded My Story Pro AI with her online business partner Wiebke — a human-centered AI storytelling coaching platform that hit five-figure MRR within 30 days of launch with zero ad spend. She now runs the business from Lisbon, where she moved with her husband and three sons after leaving Brooklyn.

She describes storytelling as “the hill I die on” — and after listening to this conversation, you’ll understand why.

Connect with Patrice through her website or LinkedIn.

About Peter Benei

Peter Benei is co-founder of the daily intelligence platform for senior marketing leaders navigating the AI era. He spent a decade as a marketing director at tech and product companies before co-founding AI-Ready CMO with .

Connect with Peter via his newsletter or his LinkedIn.

Top 10 Takeaways

1. No one cares about your product. They care about the story that connects your world to theirs.

2. Storytelling isn’t a soft skill anymore. It’s the competitive edge that code can’t replicate.

3. The line between human and AI isn’t taste — it’s lived experience. AI can structure a hero’s journey. It can’t tell you how it felt to be fired with a newborn while your husband just quit his corporate job.

4. Most people are afraid to go deep into themselves. That fear is why AI output stays shallow, regardless of how good the prompt is.

5. AI wants to make you happy immediately. A good storytelling coach knows you have to earn the depth.

6. Facts and data are now table stakes. Anyone can access enterprise-grade research in minutes. The scarce asset is emotional connection.

7. The brands that win in the next 3 years won’t be the ones with the best AI workflows. They’ll be the ones whose people actually show up as humans.

8. Storytelling compounds inside organizations. One leader who starts being vulnerable in team meetings can transform an entire unit’s culture.

9. The Stanford graduate with 105 interviews and zero offers isn’t losing because she lacks AI skills. She’s losing because she hasn’t learned to make someone feel something.

10. Your story is not copyable. The structure can be lifted. The feeling can’t.

Why no one cares about your product

There’s a version of marketing that’s been standard practice for decades. You lead with features. You cite your stats. You build a pricing page and a comparison chart and you wait for the rational decision to happen.

Patrice Poltzer watched that approach work on the TODAY Show — briefly, in the early days — and then watched it collapse.

“Marketing has changed,” she said.

The problem isn’t that founders don’t believe in their products. It’s that they treat the product and the humans behind it as separate entities. There’s the company, and then there’s the person who built it — and the assumption has been that keeping those two things apart was professional.

That assumption is wrong. And in a world of product abundance, it’s also commercially fatal.

The distinction Patrice draws is precise: storytelling isn’t about you. It’s about finding the intersection between your lived experience and the thing keeping your customer awake at 3am. When those two things meet, something neurological happens. She calls it neurocoupling. The listener’s brain synchronizes with the storyteller’s. That doesn’t happen when you list your features. It only happens when you make someone feel recognized.

Peter made the same point from the CMO side. For a decade across product-led tech companies, the instinct was always to lead with the roadmap, the integration, the metrics. The breakthrough was realizing that none of that is what closes. What closes is the moment a customer thinks: that’s me.

The problem with AI-generated stories

My Story Pro exists because of a moment of recognition.

Patrice’s students started submitting AI-generated stories. She could tell immediately. Not because the grammar was wrong or the structure was off — but because nothing in them was real.

“The problem is that most people are afraid to go deep on themselves,” she said. “And so when you have a lot of people not wanting to go deep, even if you have the best prompt loaded into Claude from the most amazing marketer, it is never going to go as deep as it should.”

The reason is structural. General AI tools are designed to give you output. They want to make you happy immediately. That’s the product. A good human interview — the kind Patrice ran under deadline pressure at the TODAY Show — works the opposite way. You earn the depth. You go through questions one through eight to psychologically disarm a person before you ask the question that actually matters.

Most AI tools skip straight to nine. And the output looks fine. But it sounds like everyone else.

This is the strategic gap My Story Pro is built to close. The platform functions less like a content generator and more like a trained interviewer — one that refuses to let you stay on the surface, because the surface is where all the slop lives.

Where the human line actually is

Six months ago, Peter would have drawn the line at taste — the human judgment required to evaluate AI output and decide what’s good enough to ship.

His answer has changed.

If taste was ever the moat, it’s getting shallow. The new line is higher up.

“The line right now is for a human to figure out the big picture for the brand. Figure out the story. What is your brand? What is the very high-level strategy? What’s the story you’re trying to tell? Once you figure that out, it’s easier to create the executional level of those stories.”

Core narrative is a human job. Execution is increasingly not.

Patrice pushed it further. The reason her product keeps growing — even as Claude, GPT, Gemini, and Grok all get better every month — is that she and her co-founder hold a fundamentally different view of what AI is for.

The problem isn’t the tools. It’s that the tools are being asked to replace the part of you that no tool can access: what actually happened to you, and how it felt.

The case for storytelling as a hard skill

The Wall Street Journal ran a piece recently: companies are desperate for storytellers. Netflix listed a Chief Storyteller role at $700,000.

This would have been unimaginable when Patrice first left the TODAY Show in 2017. “Storytelling was seen as a soft skill,” she said. “People would say, ‘Oh, storytelling — we do content marketing.’”

That framing is gone. Storytelling is now listed alongside engineering and data science as a capability gap at the enterprise level. And the reason is straightforward: everything else can be automated. The emotional vehicle — the thing that actually changes someone’s decision — cannot.

The story Patrice tells to illustrate this is about a digital media CEO in Chicago. Multi-million dollar operation. Small team. On every measurable metric, his company outperformed the competition. He walked into a meeting with a major bank and spent the first six minutes explaining who he was. No one knew his name. One person thought he was from his competitor.

He realized he was going to lose.

Not because his product was worse. Because his competitors had a face, and he didn’t. He wasn’t getting invited to the private dinners. Not being asked to speak at Cannes or South by Southwest. He was being professionally invisible while his numbers were excellent.

He started showing up differently. Sharing why the work mattered to him. What he’d been through. Patrice ran into him a year and a half later at Cannes, where he was now speaking. His team would walk through fire for him.

“That’s what we need more of,” she said. “Not a performative way. A real way.”

What this means for your team

The implication most leaders miss: this isn’t just a personal brand conversation. It’s an organizational one.

Patrice worked with a product manager at Apple — 13 years in, starting to feel the ground shift. She wanted to know whether to learn more AI, more tech. The answer was neither. She started opening her team meetings differently. Just briefly — here’s a little about my past, here’s why the way we do things matters to me.

The result was a promotion, recognition in the company's internal newsletter, and a team that completely transformed the way they worked. Her unit became cited internally as an example of what was possible.

“Who is making your brand narrative? It’s people,” Patrice said. “And you want motivated, interesting people in that mix. People who are comfortable being more human in a real way.”

The companies that win the next cycle won’t be the ones with the cleanest AI workflows. They’ll be the ones who figured out that their people are the content strategy — and gave them the language and permission to show up that way.

That’s a harder build than a prompt library. It requires actually knowing your own story.

My Story Pro is built on the idea that most people don’t — yet.

7 out of 10 marketers say they’ve received “workslop” from colleagues.

The term was coined late last year by researchers at BetterUp Labs and Stanford. It means AI-generated work that looks like it took effort but doesn’t move anything forward. The strategy deck reads like a ChatGPT default. The customer email is technically correct, but feels hollow. The campaign brief that hits every buzzword and says nothing. 27% of marketers report receiving this stuff regularly — not once or twice, but as a pattern. It damages trust within teams, forces rework, and turns AI’s promised productivity gains into a net negative.

If you manage a sales team, a support operation, or a creative department, chances are there is “workslop” in your inbox right now.

The cause isn’t hard to find: companies bought the tools and skipped the training. Only 20% of marketers say their organization’s skills development meets current needs and can adapt as new skills emerge. Twenty percent. And yet the single biggest differentiator between market leaders and followers this year wasn’t budget or tech stack or choice of AI model. It was an investment in people and skills.

The gaps are specific: 44% of organizations lack practical AI implementation skills (starting with basic prompting), 43% can’t integrate AI into existing workflows, and 37% don’t have enough technical understanding of how these tools work. This is all foundational stuff. None of it is cutting-edge.

Market leaders in the report aren’t using different tools than everyone else — same vendor pool, same platforms. They’re almost twice as likely to report positive consumer responses to their GenAI integrations, and more than twice as likely to say those integrations exceeded expectations. The difference is that they paired deployment with structured upskilling and set explicit quality standards for AI-assisted output.

That last part is worth repeating: explicit quality standards. Telling your team “use AI” without defining what acceptable AI-assisted work looks like is how you get workslop. Set the bar (yes, this means saying out loud that a raw ChatGPT dump is not a deliverable), invest in practical training on prompting and workflow integration, and connect it to business goals. Not a Slack channel with prompt tips, please. Ongoing capability building, same as you’d do for any other core skill.

We keep saying AI skills are management skills. Now there’s a 1,400-person dataset backing it up. Teach your people how to use the tools, or watch them generate workslop at scale.

— Torsten and Peter

Share with a colleague. Earn paid membership months.

Refer a friend

3 AI Marketing Tools To Try Today

SitesGPT

→ Try it here

SitesGPT uses AI to generate fully responsive websites in minutes by analyzing your business information and applying industry best practices. Use it to spin up professional landing pages and microsites without design or dev resources.

Kittl

→ Try it here

Kittl is an AI-first design platform with text-to-vector generation, AI image creation, background removal, and 26 AI models for producing logos, apparel, posters, and brand assets directly in your browser. Use it to create professional graphic designs and print-ready assets without needing Photoshop or Illustrator skills.

Lusha

→ Try it here

Lusha provides AI-powered B2B contact data, including 280M+ verified contacts, buyer-intent signals, and automated lead recommendations based on your ideal customer profile. Use it to skip the manual research grind and get accurate decision-maker emails and direct dials for outbound campaigns.

Paid members, your toolkit and this month’s report are below.

Read more

They just published V2. What was “Capable” ten months ago is now “Unacceptable.”

In the original rubric, a marketer who used AI to draft social posts and then edited them by hand was considered Capable — meeting the minimum bar. In V2, that exact behavior is listed under Unacceptable. The new Capable requires AI embedded into core work, repeatable systems rather than one-off prompts, and a clear, measurable impact on quality or efficiency. Ten months.

For marketing specifically, the new Adoptive tier — one level above the minimum — now includes running AI-driven A/B experiments with measurable results, building content systems that draft, format, and schedule across channels, and having always-on agentic workflows that operate without human involvement.

That was aspirational a year ago. Now it’s the expected trajectory for someone who wants to be seen as more than baseline competent.

If you’re a senior marketer reading this and thinking “we’re doing fine” because your team uses ChatGPT for briefs and Midjourney for mockups, Zapier just told you that’s the floor, not the ceiling. And the floor just moved above your head.

Sales wasn’t even in V1. Now it has its own rubric, and it’s not gentle.

The original framework covered engineering, product, support, marketing, and people/HR. Sales was absent. V2 adds it — along with business ops, corporate development, finance, data, and others — which tells you something about how quickly AI competence has spread from “tech team concern” to “everyone’s problem.”

The headline: if the extent of your AI usage is reading Gong summaries after calls, using AI to help write outreach emails, and using AI to find information for account plans, you are Unacceptable. Not Capable. Not “getting started.” Unacceptable. If you “cannot point to evidence that outreach is more effective, prep is faster, or pipeline has improved as a result,” you don’t meet the bar.

Capable for sales now means having built a personal prospecting workflow that saves significant time and generates pipeline, and being able to position AI transformation strategies with customers. The Adoptive tier requires chaining multiple AI tools into repeatable workflows and proactively teaching teammates what works.

But the most significant change is not even the higher bar. It’s the new fourth dimension: accountability.

V1 measured three things: AI mindset, strategy, and building. V2 adds accountability as a standalone signal. Zapier now explicitly hires for people who define what “good” looks like before they start, evaluate outputs critically, and catch what’s wrong before it ships: “With AI, you can delegate the work, but not the accountability.”

This connects to something we keep saying in this newsletter: AI skills are management skills. Most companies are still hiring for AI usage: do you use the tools, how often, which ones. Zapier is now hiring for AI judgment: do you know what a good output looks like, can you catch a bad one, and do you own the result either way? That’s a fundamentally different question, and it’s one most hiring processes aren’t set up to answer.

They’re also now looking at what they call “slope” — not just where you are on AI fluency, but how you got there and how fast you’re moving. Someone who plateaued eight months ago on the same three tools is a different candidate from someone actively experimenting and building. The signal they want is forward momentum.

You don’t have to adopt Zapier’s rubric wholesale. But you’d be wise to ask yourself: if someone applied this framework to your team today, where would they land? If the answer makes you uncomfortable, that discomfort is the point.

— Torsten and Peter

Share with a colleague. Earn paid membership months.

Refer a friend

3 AI Marketing Tools To Try Today

Rytr

→ Try it here

Rytr is a lightweight AI copywriting assistant that handles blog drafts, emails, ad copy, and social posts on a freemium model. Use it to remove the blank page problem at an unbeatable price point for solo marketers and startups on tight budgets.

Make

→ Try it here

Make is an AI-enhanced visual automation platform that connects your entire marketing stack through drag-and-drop workflows with conditional logic and API calls. Use it when your automations need branching paths, error handling, and complex multi-step sequences beyond simple triggers.

Synthesia

→ Try it here

Synthesia generates professional AI presenter videos in 140+ languages, replacing expensive video shoots with scalable, on-brand content production. Use it as the enterprise standard for AI video when compliance, custom avatars, and scale matter.

Paid members, your toolkit and this month’s report are below.

Read more

We’ve talked a lot about companies freezing junior hiring as AI takes over entry-level workflows. That was the demand side drying up. Now the supply side is following. China’s Communication University of China — one of the country’s most prestigious schools for creative talent — just suspended 16 undergraduate degree programs, including photography, comics, visual communication design, and translation. The party secretary said that the era of “human-machine collaboration” has arrived, and these programs no longer make sense as standalone degrees.

And this isn’t one university having a moment. Jilin University halted 19 arts majors in January. East China Normal University suspended 24 — including advertising and painting — late last year. Tongji University dropped visual communication and environmental design.

If you think this is some sort of a panic reaction, you are wrong. China’s universities are centrally controlled, and the Ministry of Education revoked over 1,400 programs in 2024 alone. This is policy.

And it won’t stay in China.

Western universities move slower because they’re decentralized, not because they disagree. The market logic is identical: when AI handles first-draft translation, basic image production, layout, and routine content creation, a four-year degree built primarily around those technical skills is training students for jobs that are going away. The institutions that move first just happen to be in a system where the government can push the button faster.

CUC didn’t just kill programs, though. It merged photography into film and television production. It folded comics into new AI-integrated degrees like “intelligent imaging art.” It has added 20 new majors since 2018.

The narrow, technique-heavy degree is what’s dying.

Photography as a craft still matters — a four-year program that teaches you only how to operate a camera is a different story. Same with translation: the skill is valuable, the degree that produces human translation machines competing against actual machines is not.

If you manage a creative team, this is the same story playing out in your org chart. The pure specialist — “I run Facebook ads,” “I do email design,” “I write product descriptions” — is the photography major of the corporate world.

The value is shifting toward people who can think across the stack: who understand why a campaign works, who can adapt when the tools change underneath them.

When you’re hiring juniors (if you still are), that’s the filter. Can they work with and around tools that will replace the tools they know today?

— Torsten and Peter

Share with a colleague. Earn paid membership months.

Refer a friend

3 AI Marketing Tools To Try Today

Birch

→ Try it here

Birch is an AI-powered ad automation platform that manages, optimizes, and scales paid campaigns across Meta, Google, TikTok, and Snap with real-time KPI monitoring. Use it to stop babysitting ad accounts and let automation handle bid adjustments, pausing, and scaling.

Guideless

→ Try it here

Guideless AI captures your screen workflows and automatically transforms them into narrated, branded video guides with AI-generated scripts and voiceovers. Use it to create product tutorials, onboarding walkthroughs, and support documentation in minutes without manual editing or recording retakes.

Lusha

→ Try it here

Lusha provides AI-powered B2B contact data with 280M+ verified contacts, buyer intent signals, and automated lead recommendations based on your ideal customer profile. Use it to skip the manual research grind and get accurate decision-maker emails and direct dials for outbound campaigns.

Paid members, your toolkit and this month’s report are below.

Read more

The ads appear below ChatGPT’s responses, clearly labeled, separate from the answer. OpenAI says they don’t influence outputs, and advertisers don’t get access to conversations. Over 600 advertisers are already in, with strong SMB interest. For a pilot this young, with this little distribution, that level of traction is hard to wave away.

Self-serve access opens in less than a month.

That’s when ChatGPT ads stop being a managed experiment for a select few and become a channel anyone can test. Agencies will build offerings around it, benchmarks will start forming, and the learning curve steepens for everyone who waited. We’ve seen this before with early Google Ads, early Facebook, early retail media. The teams that showed up first didn’t always spend the most — they just learned the fastest.

The inventory is interesting because of the intent behind it. People don’t scroll ChatGPT the way they scroll Instagram. They ask it to compare software, plan trips, evaluate vendors, troubleshoot problems.

That’s research-stage, high-consideration behavior, the kind of moment where a relevant ad can actually be useful. For B2B marketers especially, worth watching closely.

Nobody has published official CPM numbers yet, but the math is suggestive. Meta averages roughly $7–17 per thousand impressions depending on the vertical. Google search ads run $10–50+ for high-intent queries. ChatGPT is hitting $100M annualized from a small fraction of its U.S. user base with limited daily ad exposure — which implies an effective CPM at or above Meta’s upper range. Advertisers are clearly paying a premium for context-aware placement inside a conversational interface. Whether that premium holds as inventory scales is the open question.

Nobody should rush budget in blindly. Measurement is immature, attribution is thin, the creative playbook doesn’t exist yet. But if your products win when customers are comparing and asking follow-up questions before they buy, this channel is built for that. Map your high-intent query themes now, tighten your analytics, and decide what a successful test looks like before self-serve goes live. The worst position to be in is curious but unprepared.

— Torsten and Peter

Share with a colleague. Earn paid membership months.

Refer a friend

3 AI Marketing Tools To Try Today

BotPenguin

→ Try it here

BotPenguin is an AI chatbot platform supporting live chat, lead generation, and customer support across web, WhatsApp, and social channels. Use it as an affordable entry point for adding conversational marketing and automated lead capture to your stack.

Granola

→ Try it here

Granola uses AI to transcribe meetings in real time and deliver structured summaries with action items, decisions, and key takeaways automatically. Use it for clean, focused meeting intelligence without the feature bloat of enterprise tools.

Lindy

→ Try it here

Lindy is an AI executive assistant that triages emails, schedules meetings, updates CRMs, and joins calls on your behalf. Use it to reclaim hours from admin work and refocus on strategy and high-impact marketing activities.

Paid members, your toolkit and this month’s report are below.

Read more

Strap in.

Specialization Was Always a Fiction

I (Peter) have a particular kind of trauma from the nametags we wore at the big corporate fishing boat. The fancy titles. The carefully delineated responsibilities. We played along. We told ourselves we were special.

I spent a year in enterprise marketing before co-founding an agency with Torsten in the mid-2000s. Our clients were all enterprises. To them, we were the rock stars. The ones who “dared.” I almost printed a business card that said Peter Benei, Chief Creative Ninja. Torsten talked me out of it. I ended up as “creative director,” which, let’s be honest, sounds considerably more boring. (Unrelated, but despite all my attempts not to, I circled back to enterprise years later, then found common ground in what actually interested me: midscale and high-growth startups.)

Here’s the thing I knew even then: specialization is mostly made up.

It came from factory floors. One person, one job, one output, maximum productivity. Offices copied this model wholesale — same working hours, same skills-only focus. Education followed. We ended up with specialized business practices that translated into departments, org charts, and job titles nobody outside the building fully understood.

<your name>, Lead Performance Marketing Manager, FMCG Consumer Marketing, Digital Department, at <your favorite cookie brand>. So you do ads, right? Cool. Guess what: AI wants to have a meeting with you.

We built entire professional identities around execution layers. And technology has been quietly dismantling those layers since the word “technology” existed.

Technology Was Always Coming for the Execution

This isn’t new. It never was.

ATM vs. bank tellers. The calculator vs. the people who were the calculators. The photocopier vs. the copy clerks. Each wave, the same pattern: a technology takes the execution layer, the humans move up, and we collectively pretend the previous jobs were never really just execution.

You know the cope. “It’s not AI that will replace you, but a person who knows how to work with AI will.” It’s a reasonable-sounding argument. Until you look at the actual pattern. You know the joke, right? “It’s not the car that replaces you, but the horse that can drive.”

The difference between an HR manager and a procurement manager was never deep craft. Neither was the difference between a sales manager and a marketer, or a product manager and a project manager. What actually separated those roles was familiarity with a specific execution layer. The intention behind the work. The understanding of the context.

Strip the execution away, and those differences collapse. All that is left is people doing work in Excel. Some of them talk more about “human connection,” and others talk more about “revenue” and “ROI.”

If every previous technology wave took one layer of the cake, AI is here for the entire cake. Nothing left on the plate. And unlike every previous wave, it doesn’t respect function boundaries. Marketing, sales, HR, product, finance — the execution differences between these roles are gone.

This Is Happening Now

We don’t want to be doomers. But we’d be lying to you if we softened this.

We can’t name names — NDAs being what they are — but here’s a situation we’ve heard described from multiple directions, in almost identical terms:

A global tech company. 500 to 5,000 people. Old school. Still running kanban, scrum, and daily standups in 2026. Slow-moving. Nobody at this company would call themselves cutting-edge.

In Q1, every mid-manager — heads and directors, one level below C-suite, each managing teams of 10 to 50 people — received The News. Allocated hours to figure out how to automate their team’s tasks with AI. Expected workforce reduction: 20% by Q3, another 20% by Q4, and continue the trend until “sufficient.”

If this is Monday’s meeting at a company that still does daily standups, assume it is everywhere, or will be everywhere by the end of Q2. In execution by Q3. Felt by everyone by the end of the year.

But.

There’s another signal running in parallel, from the opposite direction.

The rise of the one-person company. Not the lifestyle solopreneur. Not the startup founder with VC ambitions. Something different: 1-to-5-person operations running at the productivity and output level of mid-sized companies. Millions in ARR. Agent swarms working round the clock. One, maybe a few, humans at the center, conducting the entire operation.

Both ends of the candle. Burning simultaneously. Same shift. Two games. All happening right now.

What Actually Survives

Six months ago, we made an argument: the barrier between human and AI is taste. Knowing what’s good enough to ship. We wrote an entire series on this — [start here if you missed it]. We have also recognized that there is a fire — we wrote about it recently. We were playing with the solution in the Human Context Window, but that wasn’t a clear picture, yet.

We were half right the whole time. Now it’s getting much clearer.

Taste still matters. But not as an aesthetic judgment. Not “is this good enough?” — because AI at current capability levels will produce above-average output on the first draft across almost every execution type. Most people can’t tell the difference. The aesthetic bar has been cleared.

The taste question has moved up a level. It’s no longer is this good enough to ship. It’s should this exist at all, and why. That’s a fundamentally different question. And far fewer people are equipped to answer it.

As Jeremy Irons says in Margin Call: ”Be first. Be smarter. Or cheat.” Cheating is out. Being the smartest is great, but hard. Being first is still available — for now.

If you’re a power user of AI today, you’re still early. The UX is complex, the tools are scattered, and the workflows require genuine curiosity to build. But we are weeks, maybe months, from this becoming the default. The secret Excel formula buried in a sub-sub-menu is becoming a hotkey everyone uses without thinking.

The window is open. It won’t stay open.

Two Paths

When it closes, there will be two kinds of people still standing in knowledge work. We’re calling them the creative and the operator. Not job titles. Orientations.

The creative is the one who owns the why. The brand narrative. The brand strategy. The story the company tells. The key messages it stands for and repeats everywhere. The decisions about what should be made, why it should exist, and what it means.

Everything else is execution and practical application of that narrative. Put your brand documents into an AI, explain the context and the intention, and you will get great output. AI is extraordinary at reacting. What it cannot do is be proactive. It cannot determine whether a task should exist. Should this be done at all? If yes, how should it be framed? That question is inherently human. It is the creative’s job. And it is not a marketing job exclusively — it applies to every function where someone has to decide what work means before anyone executes it.

The operator manages the execution. If the creative figures out the why, the operator figures out the how. The operator is the starting and ending point of the AI-driven workflow — the conductor of the orchestra, the manager of the autonomous bots and agents. Not the executor. The manager.

This requires minimal understanding of specialized craft and a deep understanding of workflows, systems, and AI orchestration. Your AI-curious copywriter will probably become the operator of all copy across all campaigns. Your designer becomes the creative head on AI-driven production. You will not need all your designers. You will need a few people who can manage AI from a designer’s perspective.

This is true for every function. Not just for marketers. Every department will have its own creatives — the leadership layer — and its own operators — the AI managers. Maybe, just maybe, there will be some craftspeople who stay, if they are really, really, really good. But would you bet your career on being one of them?

So What Do You Do

If you work for someone else, your options depend on where you sit.

If you’re in leadership, focus on the creative layer. Tech literacy matters — you won’t be able to manage a team you don’t understand. But your primary job is now the brand, the narrative, the story, the strategy. You are expected to be the creative voice on the team. If you can’t answer why does this exist, you can’t lead what executes it.

If you’re a team member, become an operator. Start now. Not next quarter. Get the courses, build the knowledge, and start building apps and microproducts with AI. Figure out how to automate your own job before someone else does it for you. Because they will. That’s not a threat — it’s a description of what’s already in motion.

If you work for yourself, you have to do both at once. The story is yours — no one else can own it, no one can fake it, and a product can be copied while a story cannot. At the same time, build the machine. If you can run a swarm of agents well, you don’t need a large team to compete with mid-sized companies. At least for now. And now matters.

What We’re Building

We’ve thought hard about what this means for AI-Ready CMO.

Our core job remains what it has always been: intelligence and guidance for marketers navigating this shift. We will go deeper on the creative layer — more brand strategy, more storytelling, more talks from those who are actually living this. Our podcast continues as the platform for those conversations. Our paid program, with tools and reports, continues to support change. And our daily signal also continues.

But we’ve hit a limit. Writing about AI workflows and agent orchestration is one thing. Showing it being built, live, is something else entirely.

So starting next week, we’re launching a live workshop series on Substack. We will build AI workflows in real time — starting with marketing operations and production, then expanding beyond marketing entirely. Because specialization is dead, and the orchestration skills that matter don’t stop at the marketing department boundary.

The first workshop is next Friday, April 3, 10:00am EST / 2:00pm GMT (you’ll get an invite). In about 30 minutes, we will set up a Claude Cowork marketing workspace from zero, run a live competitor teardown, and turn the output into a targeted campaign. All this without touching a workflow builder, a template, or a single line of code.

Since the workshop series goes beyond marketing, we’re giving it its own name: Teamless.

One-person army. A highly capable human operating a swarm of AIs to execute at 100x productivity — regardless of function, regardless of industry. The workshops are live, free, open to everyone, and built to show rather than tell.

Alongside the workshops, we’re also building Teamless as a service — a program for managers who want to transform their teams before someone transforms them out of a job. More on that soon.

The workforce changed. Most managers haven’t caught up yet. That’s the gap. Stop thinking about AI as your assistant. AI is your workforce. You are the manager. Pick your track. Both require action. Neither is optional.

— Torsten & Peter

The Teamless workshop series launches next week, on Friday, on Substack. Follow along here. If you’re a leader looking to bring this into your organization, reach out directly.

Here’s something most marketing teams are getting wrong right now.

They’ve heard about GEO (Generative Engine Optimization). They know LLMs are becoming discovery layers. They’ve seen the think-pieces about AI search changing everything. And so they’ve done what marketing organizations always do when a new channel emerges: they’ve handed it to SEO.

That’s the wrong call. And it’s going to cost a lot to miss.

GEO is not a channel. It’s not a tactic you can assign to one team and check off the list. It’s an outcome. The net result of how your entire marketing operation shows up across the internet.

ChatGPT doesn’t crawl your website and rank pages. It forms opinions. Opinions built from Reddit threads, YouTube transcripts, LinkedIn articles, G2 reviews, comparison pages, Wikipedia entries, and dozens of other signals that your SEO team doesn’t own and probably isn’t tracking.

Which means if your SEO team is the only team thinking about AI visibility, nobody is actually solving it.

But the real problem isn’t technical, but coordination. Your SEO team controls structure and discoverability. PR drives external validation and authority. Social shapes narrative and distribution. Content defines positioning and use cases. Partnerships reinforce credibility. If these aren’t aligned (if they’re not all pulling toward the same narrative story), you either don’t show up or you show up inconsistently. And inconsistency in AI responses is almost as bad as invisibility.

Most companies have no idea how they’re actually represented. They’ve never run a prompt like “What’s the best [category] tool for a mid-market B2B company?” and looked at the results systematically. They don’t know if ChatGPT is recommending them, ignoring them, or, worst case, recommending their competitors in their place.

Quick sidenote: AmIOnAI’s own research puts the Pearson correlation between SEO rankings and AI visibility at 0.36. That’s not a typo. You can rank #1 on Google and not appear once in a relevant ChatGPT conversation. The channels are decoupled in ways most marketing leaders haven’t internalized yet.

This is why GEO needs to sit at the CMO level. Not to execute — you’re not going to personally write Reddit replies — but to orchestrate. Someone needs to own the outcome across teams. Someone needs visibility into what’s actually moving the needle. Someone needs to be able to say: “We’re losing this category to a competitor in AI responses, here’s why, and here’s where we reallocate budget to fix it.”

Right now, almost nobody has that view.

What Am I on AI does

Am I on AI is an AI visibility monitoring and optimization platform built for exactly this problem. You enter your brand, it generates the long-tail prompts your actual customers are asking ChatGPT, runs them through the real ChatGPT web interface (not via API!), and produces a Visibility Score: what percentage of relevant AI conversations mention your brand.

It also tells you who’s winning in your category and, critically, why. The citation source analysis is the piece we found most useful. It maps the domains ChatGPT is actually drawing from when forming opinions about your market — Reddit, G2, TechRadar, Medium, YouTube, and Wikipedia.

That’s the intelligence you need to know where to focus effort. If Reddit has 65% influence in your category, that’s not a content strategy problem. That’s a PR and community strategy problem. Budget needs to move accordingly.

The product then generates a specific, prioritized weekly action plan, not generic advice, but recommendations tied to your actual scan results, with estimated impact percentages per action. It rescans weekly so you can see what’s working.

The data points worth knowing: one company grew ChatGPT visibility from 20% to 77% in six months by focusing on Reddit engagement. YouTube transcripts get indexed by ChatGPT — meaning the spoken content of your videos becomes citable, not just the metadata. LinkedIn articles get indexed, but regular LinkedIn posts don’t. Changes can appear within days, not months, because ChatGPT uses real-time search-based grounding.

These are not hypotheticals. They’re where the leverage is.

The CMO use case

We want to be direct about why this tool is relevant to this audience specifically.

For a marketing director or VP, Am I on AI is a channel intelligence tool. It tells you where to focus effort and why. For a CMO, it’s something more structural. It gives you a shared metric your entire marketing organization can align around — an AI Visibility Score that connects outcomes to inputs across teams.

It replaces the guesswork about which team is actually driving AI brand presence with actual attribution data. That’s the foundation for rational budget allocation in a world where the discovery layer is shifting.

If you’re managing a team and you can’t answer “What percentage of relevant ChatGPT conversations recommend us?” — that’s a gap worth closing.

Seven-day free trial, no credit card required.

Run a test on Am I On AI

About the guest

Anna Levitt is the founder of Bubble Boss Co., a company that helps leaders and organizations transform with AI and adapt to the pace of change. Anne works primarily with mid-size and larger organizations navigating AI adoption — with a focus on the human and structural layers that most transformation programmes ignore.

She spent years in corporate leadership before leaving to build on her own terms, and now runs on Substack, where she writes about AI adoption, organizational design, leadership, and staying human in a disruptive world. She completed the MIT course on organizational transformation using AI and brings both practitioner experience and a coaching approach to her work.

Connect with Anna on Substack, her website, and LinkedIn.

About the host

Peter Benei co-founded the daily intelligence platform for senior marketing leaders. Peter has been serving as a CMO, marketing leader, and consultant to high-growth B2B scaleups for the past 10+ years. He has a background in advertising, working with Fortune 500 brands.

Connect with Peter on LinkedIn, or read his newsletter.

Top 10 Takeaways

1. AI adoption stalls because of people, not tools. — The tool is the easiest part to implement. The hard part is aligning people so they see this as a value opportunity, not another disruption layered on top of everything else.

2. Most organizations have a data chaos problem they haven’t admitted yet. — Marketing, especially. Scattered data, siloed processes, undocumented workflows. AI can’t fix what hasn’t been organized first.

3. Mid-size companies are better suited for AI transformation than enterprises. — Less legacy infrastructure, more flexibility. The mistake is trying to copy enterprise-grade systems instead of designing for who they actually are.

4. Don’t become a coder. Become tech literate. — Marketers don’t need to build software. They need to understand how AI infrastructure works well enough to make informed decisions for their teams.

5. The Chief AI Officer role is non-negotiable for larger organizations. — You can’t stack AI transformation on top of existing marketing, HR, or IT roles. Someone has to own it as a full-time strategic responsibility.

6. Find the hidden talent — not the loudest voice. — The people who will drive AI adoption forward are rarely the ones with the most power in the room. They’re in the middle layer, doing the actual work.

7. AI layoffs are largely an excuse. — Most mass layoffs attributed to AI are restructuring decisions that would have happened anyway. The more important story is jobs not being created in the first place.

8. The CEO's job description is about to change fundamentally. — Fixed mindsets and old operating models won’t move the needle. Leaders need to reassess their value proposition — agility and vision are the job now.

9. Cognitive overload is quietly killing adoption. — People are already maxed out. Asking teams to adopt AI on top of everything else, without slowing the pace of change, leads to burnout, not transformation.

10. A leader without a team has nothing to lead. — Companies without people aren’t companies. The human layer isn’t a variable in the AI adoption equation. It’s the foundation of the entire thing.

5 Things Worth a CMO’s Attention

1. Why AI adoption keeps stalling — and it’s not the technology

Anne has worked with enough organizations to have a clear diagnosis: AI adoption fails at the people layer, not the technology layer. The tools get bought. The pilots get launched. Then nothing sticks.

The reason isn’t usually resistance in the abstract sense. It’s that people haven’t been given a clear answer to the question, “Where does this take me?” When employees can’t see the personal upside — not just the organizational upside — adoption stalls. It becomes another initiative layered on top of an already overloaded workday.

The second problem Anne consistently finds is data. Most marketing organizations have never been forced to document how they actually work. Processes live in people’s heads, in old email chains, in folders no one has opened in two years. AI doesn’t clean that up — it surfaces it. And when it surfaces, teams realize the infrastructure for AI adoption was never there to begin with.

For CMOs driving AI transformation, this reframes the starting question. Don’t ask “which tools should we adopt?” Ask “Are our people aligned on where this is going?” and “Is our data infrastructure actually ready?” If the answer to either is no, the tool is irrelevant.

2. The case for mid-size companies — and the trap they keep falling into

Anne makes a counterintuitive argument: mid-size companies are better positioned for AI transformation than enterprises. Less legacy infrastructure, fewer entrenched processes, and more organizational flexibility to actually redesign how work gets done.

The trap is that most mid-size companies don’t treat this as an advantage. They look at enterprise and try to replicate enterprise-grade processes — the four-week onboarding, the multi-layer approval workflows, the org chart structures that made sense at 10,000 people but create friction at 200. They hire executives from enterprise backgrounds who bring those models with them.

Anne’s argument: one size doesn’t fit all, and the companies that will navigate AI transformation best are the ones that design for who they are, not who they’re trying to become. That means being honest about which processes are actually necessary and which are just inherited behavior from a bigger playbook.

For marketing leaders in mid-size organizations, this is the permission to build differently. The fact that you don’t have enterprise infrastructure isn’t a disadvantage. It means you can design AI-ready workflows from scratch rather than trying to retrofit them into a structure never built for this.

3. Tech literacy vs. coding — what marketers actually need to develop

There’s a narrative circulating that marketers should learn to code. Anne pushes back on this directly — and so does Peter. The bar is wrong. The right bar is tech literacy.

Tech literacy means being able to evaluate a new tool, understand how it works at a functional level, assess whether it fits your team’s needs, and make informed decisions without needing a developer in the room. It does not mean being able to build something. That’s a different skill set, and it’s not the bottleneck.

The reason this distinction matters: if you set “coding” as the upskilling goal, most marketers opt out. The bar feels too high, the path is unclear, and the ROI for their specific role isn’t obvious. If you set “tech literacy” as the goal, it’s accessible, relevant, and immediately applicable to decisions they’re already making.

Anne adds a second track to this: creative development. In a world where AI handles the volume and efficiency of content production, the skills that differentiate are judgment, taste, and creative thinking. Her advice — take an art course, invest in aesthetic training, and develop the parts of your marketing capability that AI can’t replicate. These aren’t soft skills. They’re the competitive edge.

4. The organizational design question — who owns AI transformation?

One of Anne’s clearest arguments is structural: AI transformation cannot be a side responsibility. It requires someone whose full-time job is owning it.

The current reality in most organizations is that AI transformation gets assigned to whoever is most enthusiastic about it — a marketing director, an IT manager, a digitally fluent HR lead. These people have existing jobs. They have teams to manage, results to deliver, and leaders asking them questions every day. Adding “drive company-wide AI transformation” to that job description doesn’t work.

Anne’s view: the Chief AI Officer role is non-negotiable for organizations of a certain size. Not because the title matters, but because someone needs to sit at the intersection of legal, HR, risk, technology, and operations — with the authority and bandwidth to actually drive change. The role is already emerging, but the job descriptions are still confused bundles of responsibilities borrowed from other functions. They’ll clarify. Companies that create this role intentionally now will be ahead of those that wait.

For CMOs, this has a practical implication. If your organization doesn’t have someone in this role, you are probably absorbing parts of it by default. That’s worth naming explicitly — both to protect your own capacity and to make the case for a dedicated resource.

5. The burnout problem that no one is putting in the AI adoption model

Anne’s most underrated point is about cognitive load. Organizations are asking their people to adopt AI, learn new tools, redesign workflows, and deliver on existing targets — simultaneously, at pace, in an environment where something new drops every week.

The human nervous system isn’t built for perpetual high-speed adaptation. Anne is blunt about this. Burnout isn’t a fringe concern in the AI adoption conversation — it’s a central variable that almost never appears in the model. Demotivated, cognitively overloaded teams don’t adopt new technologies effectively. They comply minimally and wait for the next initiative to replace it.

Her prescription isn’t a productivity framework. It’s a design question for leaders: are you pacing this in a way that humans can actually absorb? Are you giving your teams permission to go deep on one thing instead of chasing every new capability? Are you building in the kind of recovery time that sprint-based work requires?

The companies that will get AI adoption right won’t be the ones that moved fastest. They’ll be the ones that moved intentionally — that understood their people’s capacity as a real constraint, not just an HR talking point. Anne’s argument, ultimately, is that staying human through this transition isn’t a soft goal. It’s the whole ballgame.

A similar change is happening again, right now. It is even more disruptive, and I believe it will be harder for many to adjust to.

I am not talking about AI in general. Chatting with a model, as most people do, is just the bottom rung of this ladder. I am talking about agents and, more specifically, about desktop agents. They are applications that have access to and can work with local files on your computer. If you need a quarterly analysis of ad performance, you just point the agent to the relevant Excel files and it will spit out an interactive, fancy dashboard in two minutes. If you need to prepare for a call, you simply ask the app to write a brief for you based on the last three call transcripts and your sales deck. You don’t need to know how to code, how to read code, or know that code exists at all.

A desktop agent is not something you chat with. It is something that does the work for you.

The first one, Claude Cowork, was released just 10 weeks ago. We covered it, saying “if this is not AGI — basically superhuman intelligence — then I don’t know what that word means anymore.” OpenClaw, the first always-on general-purpose agent, went live two weeks later, and we said “when something captures that much attention that quickly, it usually means something is shifting underneath. This is the kind of story that will be on the front page of the Wall Street Journal by the weekend.” (Guess what, we were right.)

We talk to a lot of senior leaders, both in marketing and in the wider corporate world, and everyone is saying the same thing: they want their teams to start using desktop agents right now. Not next year, not in H2, but today.

I’ll be as blunt as possible here: we see this as the “Excel of 2026”. Lotus 1-2-3, the spreadsheet that defined the PC era, came out in 1983. Excel for Windows, four years later. By 1988, anyone not using them at work was unemployable. The productivity gains from moving away from paper ledger sheets were so obvious that spreadsheet software literacy became mandatory.

Working with agents is not something you can learn from a book. It is a skill you learn as you do it. If you have just half an hour this week, try any of these options:

Claude Cowork

The first desktop agent. Requires a $20 / month subscription ($17 if paid annually), now available both for Mac and Windows. The strongest overall performer offering incredible flexibility and smarts.

Download it here

Manus

Acquired by Meta a few months ago, Manus is a fantastic agent for marketers with built-in Meta Ads dashboard integration and Instagram content management tools. The cheapest paid plan is $34 per month.

Download it here

MiniMax Agent

A capable alternative from one of the fastest-moving teams in AI. It costs $16 per month on the cheapest plan, and you get some free credits to try the app first.

Download it here

Accio Work

A brand new desktop agent from Alibaba, just released yesterday. It targets SMEs with features like a built-in Shopify store manager. Completely free for the next two weeks. While it uses a less capable model, in some respects it is the most advanced. It lets you “employ” multiple agents with their own areas of responsibility, just like you would build your team at a company. You can even write their job descriptions!

Download it here

Cloud agents

These agents don’t live on your computer; they are hosted online instead. The main advantage, of course, is that they are always-on. The biggest drawback is that they don’t automatically have access to your files unless you upload them manually.

• Perplexity Computer is probably the most advanced cloud agent, exceptionally good at research tasks. $17 / month for the Pro plan.

• Twin is positioning itself as a “company builder” agent with skills automating replies to inbound leads and monitoring the social media accounts of competitors. $20 per month, but they have a generous free trial.

• Lindy is an agent focused on personal assistant tasks. You can even talk to her on iMessage! However, it costs $49 per month.

• Kimi Claw, built on Openclaw. Using it requires a bit more technical skills, and it is not cheap (you need the $31 / month plan). We fully expect Kimi to release a more user-friendly desktop app soon.

Yes, today’s newsletter was a bit different. No “one thing you need to know in AI today” — but we feel so strongly about this topic, we wanted to make sure we put it on your radar.

Go and try one or two agents this week. Seriously—your career might depend on it.

— Torsten and Peter

Share with a colleague. Earn paid membership months.

Refer a friend

3 AI Marketing Tools To Try Today

Adwisely

→ Try it here

Adwisely uses AI to automate Facebook and Google ad buying, optimizing bids, audiences, and budget allocation in real time. Use it to replace the need for a dedicated media buyer and let the algorithm handle campaign optimization.

Crayo

→ Try it here

Crayo is an AI short-form video creator that generates clips with auto-captions, backgrounds, and music optimized for TikTok, Reels, and Shorts. Use it to produce scroll-stopping vertical video at scale without touching a timeline editor.

Fathom

→ Try it here

Fathom is a free AI notetaker that records, transcribes, and summarizes Zoom, Google Meet, and Teams meetings automatically. Use it to capture every meeting with action items so nothing falls through the cracks.

Paid members, your toolkit and this month’s report are below.

Read more

Our proprietary tool, ScoutMap analyzes your URL and surfaces the audience segments, channels, and opportunities your current strategy is missing. Free to use. AI-Ready CMO paid members unlock the full report and project workspace.

→ Try ScoutMap for free

Ask an LLM for strategic advice about your business, and you’ll get a confident, well-structured answer that sounds like it was written specifically for you.

It wasn’t.

A new study published in HBR tested seven leading LLMs—including ChatGPT, Claude, Gemini, Grok, and others—across seven fundamental business tensions where leaders must pick a side: differentiation vs. cost leadership, long-term vs. short-term, exploration vs. exploitation, and so on. They ran over 30,000 simulations.

Guess what? Across nearly every tension, every model clustered toward the same trendy option. Differentiate, don’t commoditize. Augment, don’t automate. Think long-term, always. The researchers coined a term for this: “trendslop”—the tendency for AI to recommend what’s buzzy over what’s right for a specific situation.

The obvious fix of writing better prompts or providing richer context didn’t work. The researchers tried everything: reversing option order, changing framing, demanding a pros-and-cons analysis, and even offering rewards. Across 15,000 additional trials on ChatGPT alone, the biases barely moved. The LLM sounded perfectly tailored to each situation while quietly steering each toward the same cluster of fashionable ideas.

As the researchers put it, LLMs are like the person in the room who can articulate every buzzword from TED Talks and conference seminars but has never analyzed how these ideas play out in the messy realities of a specific market.

It’s not hard to see the marketing parallels.

Should we invest in brand or performance? Go premium or compete on value? Build something new or optimize what’s working? An LLM will almost always nudge you toward the sexier option: the brand play, the premium positioning, the innovative new channel. It will rarely tell you that your best move is to grind on operational efficiency or double down on what’s already converting. But Walmart, Costco, and many others have built empires on exactly this. Sometimes boring is the strategy.

This doesn’t mean you should stop using LLMs for strategic thinking, but it does mean you should use them differently. They’re excellent at generating alternatives, surfacing blind spots, and stress-testing a direction you’ve already chosen. They’re terrible at making the hard trade-offs that strategy actually requires.

Strategy isn’t about finding the most inspiring option—it’s about deciding what you will not do. That part stays human.

Next time an LLM gives you a recommendation that sounds smart and feels right, ask yourself: would it give this same advice to my competitor? The answer is almost certainly yes.

— Torsten and Peter

Share with a colleague. Earn paid membership months.

Refer a friend

3 AI Marketing Tools To Try Today

Make

→ Try it here

Make is an AI-enhanced visual automation platform that connects your entire marketing stack through drag-and-drop workflows with conditional logic and API calls. Use it when your automations need branching paths, error handling, and complex multi-step sequences beyond simple triggers.

Zebracat

→ Try it here

Zebracat is an AI video tool that transforms blog posts, scripts, or prompts into short-form videos with stock footage, music, and captions. Use it to repurpose written content into social-ready video without any editing skills.

Shield

→ Try it here

Shield is an AI-powered LinkedIn analytics platform that tracks post performance, audience growth, engagement trends, and content ROI across personal and company pages. Use it to surface what’s actually driving reach on LinkedIn, so you double down on what works instead of guessing.

Paid members, your toolkit and this month’s report are below.

Read more

Lau was Anthropic’s first growth marketing hire and managed paid search, paid social, app store optimization, email, and SEO for nearly ten months all by himself.

Let’s get the facts straight first, because the internet already mangled this one. Anthropic — a company now valued at $380 billion — does have brand, product marketing, and comms teams. Their Super Bowl ad wasn’t a one-man job. But the performance marketing engine — the part that’s measurable, iterative, and directly tied to growth — was.

As you surely know, growth marketing at a company of this size is usually handled by a team of a dozen people.

So, how did Lau do it? Claude Code, without any programming background (he literally had to Google how to open the Terminal app on his Mac, as he admitted this in his interview video on YouTube). What he discovered was that he could describe a problem in plain English and have Claude Code build working software for him. Not prototypes or mockups but actual tools he used in production, every day. Within a week of installing it — out of sheer curiosity, after a colleague shared a setup guide on Slack — he had built two automated workflows that changed how he operated.

First, Lau managed high volumes of visual ad assets, and creating copy variants meant manually duplicating frames, switching between Google Docs and Figma, and copy-pasting headlines one by one across multiple aspect ratios. Half an hour per batch, easily. His plugin, written by Claude Code, does it in 30 seconds.

The second workflow automated Google Ads copy generation using two specialized sub-agents — one for headlines (30-character cap) and one for descriptions (90-character cap) — because a focused agent with a single job and strict constraints will outperform a general-purpose prompt trying to do everything. He also wired up a Meta Ads integration that lets him analyze campaign performance directly inside Claude, no dashboard required. Ad creation went from two hours to fifteen minutes. Creative output went up tenfold.

None of this required him to write code. Indeed, he did not write a single line of it. He described problems in English and iterated.

The approach is almost disappointingly methodical. Identify the repetitive parts of your workflow. Check if they have API access (or ask Claude to check it for you). Break them into focused sub-tasks. Automate each one separately. Most of us skip that first step entirely — we don’t sit down and actually deconstruct how we work before trying to improve it. Lau did, and that turned out to be the real game-changer, not the AI itself.

Our job isn’t shrinking. It’s shapeshifting. The hours Lau saved on copy-pasting headlines and counting characters got redirected to strategy, creative judgment, and running more experiments than most full teams manage in a quarter.

He still reviews every piece of generated copy personally. Is the value proposition clear? Does the tone match? Does it differentiate from competitors? The AI doesn’t answer those — you do. But the mechanical grunt work is a solved problem for anyone willing to invest a few days in setup.

— Torsten and Peter

Share with a colleague. Earn paid membership months.

Refer a friend

3 AI Marketing Tools To Try Today

Publer

→ Try it here

Publer is an AI-powered social media scheduling platform with built-in content generation, link-in-bio pages, and multi-workspace management. Use it as the middle ground between Buffer’s simplicity and Hootsuite’s complexity with bulk scheduling and AI image generation built in.

Browse AI

→ Try it here

Browse AI uses AI-powered robots to scrape, extract, and monitor data from any website with no coding required and automatic change detection. Use it to automate competitive intelligence, price monitoring, and market research without building custom scrapers.

Colossyan

→ Try it here

Colossyan is an AI video creation platform that turns text scripts into professional presenter-led videos using lifelike digital avatars in 70+ languages. Use it to eliminate the need for cameras, actors, or studio time when producing training content and explainer videos at scale.

Paid members, your toolkit and this month’s report are below.

Read more

And as it spreads everywhere, something clarifies. Actually, two things.

One, we are starting to understand the difference between knowing and understanding. Despite the opportunity of knowing everything, some of us use this chance to do stupid f—ing shit. But that’s OK, we are humans. In the early 2000s, we also thought social media would bring free speech everywhere.

Two, we will rediscover the offline world. In this weekend edition, we won’t be judging everyone on how they use the available digital tools at their fingertips. But we will discuss what happens if digital simply takes care of itself, so we can go back to doing our own things. Like, I don’t know, gardening.

So let’s talk analog and the new trend of offline marketing.

We never left. But we have arrived.

We had offline. Then we built online, and for more than three decades, online ate everything. Attention, commerce, identity, relationships, and entertainment. The digital layer expanded until it felt like the whole world.

But it was never the whole world. It was an interface. And interfaces, when they mature, disappear. Every interface on this planet that is used globally has one single faith: they disappear and become infrastructure. And normally, you don’t care about infrastructure. You care only if there is a problem with it.

What we are entering is not a return to offline. That word carries the wrong implication, that digital is the default, and we are retreating from it. The opposite is true. Digital is becoming infrastructure.

Like electricity, like plumbing, like roads, it will run everything, invisibly, underneath. Governance, commerce, finance, logistics, communication, all of it will live on the digital layer, mostly handled by AI, seamlessly, without requiring your conscious attention.

No more candlemakers. No more oil lamp manufacturers. Just a couple of people running reactors producing electricity, quietly humming in the background so you can flip your switch in the bedroom.

No more information packagers, excel-junkies, PowerPoint deck builders, online traders, code writers, and oh, thanks God, no more personal brand experts for social media content. Just a couple of people running AI agents, producing results, and quietly humming in the background so you can radically decrease your screen time.

What remains above that layer is analog. Not as a consolation. As a destination. This is the Analog Market. And it is where most of your customers are going.

Looking back, it was obvious.

The signals were visible in 2025. They were being systematically underread by marketing teams too deep in their digital funnels to notice.

Forrester predicted that a third of consumers would choose offline over online brand experiences. Harris Poll found 77% of Gen Z and Millennials had planned travel or outings specifically around visiting a physical store, not out of necessity, but out of desire. 73% said shopping at a pop-up felt like being part of a cultural moment. Edelman’s Trust Barometer found only 42% of people trusted social media as an information source. Europol forecast 90% of online content will be AI-generated by 2026.

When the entire digital layer becomes indistinguishable and authorless, the physical object becomes the only thing with a clear origin. The only thing clearly made, by someone, for someone.

Two philosophers had mapped the intellectual territory before the data caught up. Byung-Chul Han argued in 2022 that the digital world had become a world of non-things, content with no weight, no origin, no soul.

Physical objects accumulate history. They carry an aura. When the infosphere tips toward content so generated that it has no author, the physical object becomes the exception. Which makes it precious.

Moeller and D’Ambrosio, in You and Your Profile, identified the trap built into digital identity, what they called profilicity: identity constructed through how we are observed by others, endlessly calibrated for a general audience. It worked as long as there was a signal in the profile. But when AI generates infinite, perfect digital profiles, differentiation collapses.

The only thing that cannot be faked or optimized is physical experience. Where you go. What you touch. What you show up for. Analog becomes the last unfalsifiable signal for humans, consumers, and brands.

Two markets. Two games.

Here is what most marketing frameworks are not yet accounting for: digital and analog are not just two channels. They are two distinct markets, with different buyers, different decision-making mechanisms, and, soon, fundamentally different marketing logic.

The digital market is becoming the infrastructure market. Commodities, utilities, information products, subscriptions, logistics. Everything where price, speed, and convenience are the decision criteria.

This market will increasingly be marketed to, purchased by, and optimized by AI. Your customer is not browsing your digital ad. Their AI agent is evaluating your product against seventeen alternatives and making a recommendation. The human may not be involved in the transaction at all.

Marketing here becomes machine-legible: structured data, clear value propositions, competitive pricing, frictionless purchase flows. You are not persuading a person. You are satisfying a query.

The analog market is something else entirely. This is where premium, luxury, and upscale lifestyle brands live, and increasingly, exclusively. It is the market for things with weight, craft, history, and soul.

Things that require a human to encounter, consider, and decide. Things where the experience of acquiring is part of the value of owning. This market cannot be marketed to by AI because the purchase is not a transaction. It is a moment. It is human-marketed, for humans only. And it commands the margins that the digital infrastructure market will never see again.

We all know this: diamond prices are not high because there are few diamonds. They are high because traders keep them locked up to inflate the prices. A luxury watch is expensive, not because it cannot be made on an assembly line. But because only a few are made by hand for each model.

When access to information is universal and free, you have to manufacture scarcity. When digital is endless and abundant, analog must be the ultimate scarcity.

The bifurcation is already happening. Mass consumer goods are racing toward AI-optimized digital purchase flows. Meanwhile, the brands that built meaning in physical space, through stores worth visiting, events worth attending, and objects worth keeping, are pulling away on brand equity, customer loyalty, and price premium.

Not because they rejected digital. Because they understood that their customer, at the moment of decision, is analog. Their customer is not just making a transaction. It’s creating and living an experience.

Analog market = analog marketing

If the analog layer is where human experience lives, then the marketing discipline that defines this decade is experience design. The craft of building encounters that are worth having, and worth remembering.

Print. Not nostalgia. Scarcity. In a world of frictionless AI-generated content, a beautifully produced physical catalog is remarkable precisely because it required intention, cost, and craft. It sits on a table. It gets kept. In 2026, 79% of Millennials said they looked forward to receiving brand catalogs. 64% of Gen Z used them as room decor. Harvard Business Review data showed print catalogs generating an average of $850 in annual purchases per consumer. Print is not dying. It is becoming premium.

Events and conferences. When digital presence is automated and ambient, showing up is a statement. The conference, the meetup, and the branded gathering are not logistics events. They are the primary arenas where trust is built, and decisions are made. The CMO investing in a room of real people will outcompete the one optimizing a funnel nobody is looking at.

Read more